# Replicate our approach of using dictionary features to predict income with a 
# different dataset from (Pennebaker et al., 2014 [full citation in main text])

plos_df <- read.csv("plos_data.csv", header = TRUE)
plos_df <- plos_df[!is.na(plos_df$sat_equiv), ]
summary(lm(sat_equiv ~ ., data = plos_df))
